Check Clearing, Payment Systems, and Float#

Here’s something that sounds impossible: under the right conditions, money can exist in two places at the same time.

Not metaphorically. Not as an accounting trick that gets corrected before anyone notices. For real — at least temporarily. The banking system, for a brief window, genuinely contains more reserves than it should. Both the sender and the receiver have the money. The books say so. The numbers add up to more than the total.

This phenomenon has a name: float. And for decades, it was one of the most powerful — and strangest — disturbances to bank reserves in the entire financial system.

How a Check Splits Money in Two#

The mechanics are easiest to see with a paper check, where the time gaps were biggest.

Say a customer at Bank A writes a $5,000 check to a supplier who banks at Bank B. The supplier deposits it. Bank B, following standard procedure, credits the supplier’s account within a day or two — the money shows up as available. But Bank A hasn’t heard about any of this yet. The check has to physically travel from Bank B to a Federal Reserve processing center, get sorted, and land on Bank A’s desk.

During that transit — which used to take one to three business days — a strange situation exists. The supplier’s account at Bank B shows $5,000 more. The customer’s account at Bank A hasn’t been touched. Both banks carry the $5,000 on their books. The money, for the moment, exists twice.

At the reserve level, the same doubling happens. The Fed credits Bank B’s reserve account when it processes the deposited check. But it hasn’t debited Bank A yet, because the check hasn’t finished its journey. Total reserves in the system have temporarily increased by $5,000 — not through lending, not through open market operations, but through a processing delay.

That temporary bump is Federal Reserve float.

This Used to Be a Big Deal#

In the paper-check era, float wasn’t some footnote. It was a force.

At its peak in the early 1980s, Federal Reserve float regularly topped $6 billion a day. When bad weather snarled mail delivery or processing centers got backed up — a good Northeastern snowstorm could do the trick — float spiked to $10 billion or more.

To put that in perspective: total reserves in the system at the time ran around $40 to $50 billion. A $6-to-$10-billion float spike meant reserves were temporarily inflated by 12% to 25%. That phantom injection pushed down the federal funds rate, eased interbank lending, and effectively loosened monetary policy — without the Fed lifting a finger.

The Fed’s open market desk had to scramble to counteract these surges. High float? Conduct reverse repos to drain the excess. Low float? Add reserves. Managing float became a daily chore for the monetary policy team, and a frustrating one at that, since it depended on weather forecasts as much as financial analysis.

Schrödinger’s Dollar#

There’s an analogy that captures this perfectly: quantum superposition. In physics, a particle can exist in multiple states simultaneously until it’s observed, at which point it collapses into one definite state.

That $5,000 check in transit works the same way. Until clearing is complete, the money occupies a superposition — simultaneously in the sender’s account (not yet debited) and the receiver’s account (already credited). The “observation” — the moment the check finally clears — collapses the superposition. The sender’s account gets debited, Bank A’s reserves drop, and the doubling resolves.

This isn’t just a cute comparison. It gets at something real about the nature of money. Money isn’t a physical thing that can only be in one location. It’s an accounting entry — a record of who owes what to whom. When the accounting system processes claims at different speeds for the two sides of a transaction, the total of all claims can briefly exceed the total of all actual reserves. The system runs in a state that, if frozen permanently, would be impossible. But because the superposition always resolves — the check always clears — it holds together.

From Days to Milliseconds: The History of Shrinking Float#

The story of float is really a story about processing speed. Every technology upgrade compressed the gap, and the gap is what created float.

The Paper Era (pre-1970s): Checks traveled by mail. Sorting happened by hand. Cross-country payments took three to five days. Float was enormous, variable, and hostage to the postal service and the weather.

The Automated Era (1970s–2000s): Automated clearinghouses (ACH), electronic imaging, and high-speed sorting equipment steadily chipped away at transit times. The Monetary Control Act of 1980 forced the Fed to price its check-clearing at market rates, pushing efficiency further. Float started its long decline.

The Check 21 Era (2004–present): This was the big one. The Check Clearing for the 21st Century Act let banks process digital images of checks instead of shipping the physical paper. A check deposited in Los Angeles could clear against a New York bank in hours, not days. Float collapsed. By the late 2000s, it was typically under $1 billion on most days — a rounding error compared to its former glory.

The Real-Time Era (2023–present): The FedNow Service, launched in July 2023, enables real-time settlement for everyday payments. Sender debited, recipient credited, simultaneously — or within seconds. When there’s no time gap, there’s no float. The phenomenon effectively ceases to exist.

What Float’s Disappearance Means#

The near-elimination of float is one of the cleanest wins of technology over a systemic disturbance. A factor that once caused billions in daily reserve swings and demanded constant Fed attention has been engineered into irrelevance.

Several things follow from this.

First, the Fed’s daily operations got simpler. The open market desk no longer needs to estimate float each morning and figure out how to offset it. One major source of unpredictable volatility just went away.

Second, reserve accounting got more honest. Reserves now reflect genuine deposits and central bank operations rather than being padded by processing delays. The gap between “what the numbers say” and “what’s actually there” has largely closed.

Third, the banking system lost a hidden subsidy. Float gave banks collective access to reserves they hadn’t earned through deposits or borrowing. That phantom liquidity supported extra lending and eased settlement. Its disappearance tightened the relationship between actual reserve holdings and banking activity.

Float Beyond Checks#

While check float has largely vanished, the concept shows up in other payment channels — though at much smaller scales and shorter durations.

Wire transfers via Fedwire settle in real time. No gap, no float.

ACH transactions — payroll deposits, bill payments, person-to-person transfers — typically settle next-day or same-day. There’s some residual float, but it’s small and predictable. The expansion of same-day ACH processing in 2018 squeezed it further.

International payments through SWIFT and correspondent banking can still involve multi-day delays, creating float across borders. But that’s a cross-jurisdictional phenomenon involving different central banks — a different beast from domestic Fed float.

Credit card settlements have their own one-to-three-day lag for merchants, but this runs through card network settlement systems rather than Fed reserve accounts, so it doesn’t directly inflate aggregate reserves.

Float’s Place in the Reserve Map#

In the framework mapping forces that act on reserves, float holds a unique spot. Cash withdrawals and government transactions represent genuine movements — reserves flowing between the banking system and external actors. Float is different. It’s a timing artifact. It doesn’t move reserves from one sector to another. It temporarily inflates the measured level by exploiting a gap in the accounting process.

That makes it a phantom variable — something that shows up as real in the data but has no underlying economic substance. The reserves float creates can’t be permanently deployed. They exist only until clearing catches up and the superposition collapses. Yet during their brief life, they’re indistinguishable from the real thing. Banks used float-inflated reserves for overnight lending in the federal funds market without hesitation.

The disappearance of float demonstrates something broader: infrastructure modernization can eliminate entire categories of monetary disturbance. The forces acting on reserves aren’t fixed by nature. They’re products of institutional design and technology. As payment systems evolve, the map of what matters has to be redrawn.

We’ve now looked at individual behavior (cash withdrawals), government activity (taxes and spending), and systemic infrastructure (float) as reserve disturbances. The next article turns to another factor — the fees and charges banks impose on their customers — and how these seemingly minor revenue items create their own subtle effects on the reserve landscape.

The Bigger Lesson#

Float’s story is, at its core, a story about obsolescence. Something that once demanded daily attention from the Fed’s trading desk has been compressed into nothing by digital processing and real-time settlement. The paper check — once the backbone of American payments — became a relic, and with it, the monetary magic that briefly let money exist in two places at once.

But the deeper lesson outlasts the specific phenomenon. The factors affecting bank reserves aren’t permanent features of the financial landscape. They’re contingent on technology, regulation, and institutional practice. What matters today might not matter tomorrow.

Float was contingent. The time gap that created it was an artifact of paper and mail. Once technology closed the gap, the phenomenon vanished. Other disturbances — cash demand, government fiscal flows — are more structural, but even they evolve as economies and institutions change.

The map of reserves is never finished. It has to be redrawn as the terrain shifts beneath it.